linux-kernel - Re: Commit 6016af "[media] v4l2: use __u32 rather than enums in ioctl() structs" breaks C++ users of V4L2

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [day] [month] [year] [list]

Date:	Tue, 17 Jul 2012 07:39:35 +0300
From:	"Rémi Denis-Courmont" <remi@...lab.net>
To:	"Jason L Tibbitts III" <tibbs@...h.uh.edu>
Cc:	Sakari Ailus <sakari.ailus@....fi>,
	Mauro Carvalho Chehab <mchehab@...hat.com>,
	Hans Verkuil <hans.verkuil@...co.com>,
	linux-kernel@...r.kernel.org
Subject: Re: Commit 6016af "[media] v4l2: use __u32 rather than enums in ioctl() structs" breaks C++ users of V4L2

Le lundi 16 juillet 2012 23:40:01 Jason L Tibbitts III, vous avez écrit :
> I ran into problems compiling the program ZoneMinder on Fedora rawhide
> (currently using something around 3.5rc6) which do not appear with 3.4
> kernels.  With help this was traced to commit
> 6016af82eafcb6e086a8f2a2197b46029a843d68, "[media] v4l2: use __u32
> rather than enums in ioctl() structs" which changed videodev2.h in a way
> which appears to be incompatible with C++.
> 
> This results in code such as the following:
>         enum v4l2_buf_type type = v4l2_data.fmt.type;

That code is dangerous, probably even more so in C++. If the running kernel 
version is more recent than the kernel header version that userland was 
compiled with, the field value could be outside the range of the enumeration.

> failing to compile with:
>   zm_local_camera.cpp:1523:49: error: invalid conversion from '__u32
>   {aka unsigned int}' to 'v4l2_buf_type' [-fpermissive]
> but only when compiled with the headers from a 3.5 kernel.
> 
> I'm very far from a C++ expert.  I talked with some people who do grok
> it and the issue comes down to restrictions on assignments of ints to
> enums


> and additionally that enums in C++ don't have defined size.

That cannot be true. Your C++ compiler must agree with the kernel's C compiler 
on the size of enum. If they did not, V4L2 would not have ever worked from C++ 
code in previous kernel versions.

-- 
Rémi Denis-Courmont
http://www.remlab.net/
http://fi.linkedin.com/in/remidenis
--
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@...r.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/